The ubiquitous coverage, which is an ability of communication terminal devices such as mobile phone terminals communicating by wireless anywhere in the world has not been actualized today, but has been under development.
Their mobile systems include GSM, GPRS, EDGE, WCDMA, DCS and PCS cellular systems. The systems have the feature of signals according to a constant envelope and the feature of signals following a changing envelope. In addition, there have been growing demands for multiband and multimode materialized by a wider range of combinations of the time division multiplex and code division multiplex. Now, it is noted that GSM is an abbreviation of Global System for Mobile Communication, and GPRS is an abbreviation of General Packet Radio Service. EDGE is an abbreviation of Enhanced Data for GSM Evolution or Enhanced Data for GPRS. WCDMA is an abbreviation of Wideband Code Division Multiple Access. DCS is an abbreviation of Digital Cellular System. PCS is an abbreviation of Personal Communication System.
There is a description about a typical two-step transmitter in a non-patent document presented by Abdellatif Bellaouar, “RF Transmitter Architectures for Integrated Wireless Transceivers”, The Eleventh International Conference on Microelectronics, 1999, 22-24 Nov. 1999, pp. 25-30 (hereinafter referred to as “Non-Patent Document 1”). The typical two-step transmitter has a quadrature modulator including mixers, a π/2-phase divider and an adder, a first band-pass filter, an RF mixer, a buffer amplifier and a second band-pass filter, in which an output signal of the second band-pass filter is supplied to an RF power amplifier. In such two-step transmitter, a pair of base band signals I and Q is supplied to one input terminals of the two mixers of the quadrature modulator, an intermediate-frequency local signal is supplied to an input terminal of the π/2-phase divider, and two outputs of the π/2-phase divider, which differ in phase by π/2 (90°), are supplied to the other input terminals of the two mixers. Two output signals of the two mixers are supplied to two input terminals of the adder. Thus, a baseband signal is converted up to an intermediate frequency of e.g. 70 MHz by use of the intermediate-frequency local signal. Between an output of the adder of the quadrature modulator and one input terminal of the RF mixer is connected a first band-pass filter for removing harmonics of the intermediate frequency. To the other input terminal of the RF mixer, an RF (radio frequency) local signal is supplied. An RF output signal from the RF mixer is amplified by the buffer amplifier, and then supplied to the second band-pass filter for removing an undesired sideband. The solution of using a filter for the purpose of attenuating a high-level sideband would be very simple and work well with a low electric power, however such filter is hard to realize and needs a physically large off-chip size. Further, Non-patent Document 1 presented by Abdellatif Bellaouar introduces a direct up-conversion architecture which can be realized with a smaller number of devices, too. According to the architecture, a baseband signal is directly converted into an RF transmit signal at the output of the adder of the quadrature modulator by use of an RF local signal supplied to the π/2-phase divider of the quadrature modulator.
The direct up-conversion (DUC) architecture is introduced as a promising art which can materialize a high-packing density transmitter by decreasing costly external parts, in another non-patent document presented by Gabriel Brenna et al, “A 2-GHz Carrier Leakage Calibrated Direct-Conversion WCDMA Transmitter in 0.13-μm CMOS”, IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 39, NO. 8, AUGUST 2004, PP. 1253-1262 (hereinafter referred to as “Non-patent Document 2”). According to the DUC transmitter architecture, I and Q transmit baseband signals are supplied to I/Q modulator through corresponding baseband filters. The I/Q modulator converts transmit baseband signals to a radio frequency (RF). With the radio frequency, I and Q signals are combined, and amplified. The signals undergo external filtering and further amplification, and are supplied to a duplexer before they are sent out trough an antenna. To reduce the oscillator pulling by the output of a power amplifier (PA), the local oscillator (LO) is set to 4 GHz, which is twice a carrier frequency. For the purpose of producing an accurate quadrature local signal of 2 GHz, a digital divider is used. Also, in this document, it is reported that carrier leakage is a serious drawback of DUC architecture. The carrier leakage forms an interference signal of about 1.9 GHz within the WCDMA signals' frequency band ranging 1.895 to 1.905 GHz. The carrier leakage will cause EVM (error vector magnitude) and ACPR (Adjacent Channel Power Ratio) to deviate from the specifications. Further, according to this document, offset calibration in which a 6-bit current source is used for two operational amplifiers of baseband filters is adopted to suppress the carrier leakage. Still further, according to this document, carrier leakage calibration in which a 5-bit binary weighting current source is used is adopted to suppress the carrier leakage. The power of carrier leakage is detected by an on-chip power detector in the condition of no transmit signal. The resultant analog output voltage of the detector is converted into a digital signal by use of an automatic digital calibration algorithm. The algorithm controls the modulators and a calibration circuit for the baseband filter so that the measured carrier leakage is minimized.
The patent document JP-A-2003-152558 contains the description about prevention of the interference owing to the entry of a modulated transmit signal into a local oscillator in the case where the transmitter frequency agrees with the frequency of the local oscillator in the direct conversion system. Therefore, according to the document JP-A-2003-152558, a local signal from the local oscillator is supplied to a band-pass filter, whereby a predetermined harmonic (e.g. third harmonic) is eliminated. The harmonic is divided with a predetermined dividing factor to make setting so that the transmit signal frequency is not an integer multiple of the number of the oscillating frequency of the local oscillator.
In the patent document JP-A-2003-324366, it is described that the ratio C/D of the receive frequency vs. the oscillating frequency of the receiver-system local oscillator is made different from the ratio A/B of the transmitter frequency vs. the oscillating frequency of the transmitter-system local oscillator in order to reduce the bad influence on the receive sensitivity by the interference by receiver-system and transmitter-system local oscillators.